My signal pentodes misbehaving, make 6BR7&5879 sound bet

[chopstuck]

I have built a few amps designed for the EF86 as a lone pre stage and as a second boost stage. I have tried them with both the 6BR7 and the 5879.
At first I thought they sounded OK but after just a little while they became grating and two dimensional.

Now just changing the socket wiring without changing values and subbing the tubes works but not well enough. I notice big changes in EF86 applications when using a 6BR7. Mostly too much drive, plate and screen voltages are way off and current draw has increased. I tried a 5879 in a EF86 special. It worked better than a 6BR7 but not as well as the EF 86.

In both AC15 and the EF 86 special designs I spent time juggling plate and screen resistor values to get the voltages in spec with published ones. This made the meter happy but it didn't improve the amp at all. Wrong avenue.

I have the tube curves in front of me and and understand the most basic aspect of the traces . Could someone help me find a way of getting similar
performance from these more affordable pentodes in relation to setting bias and plate and screen voltage and current.

[martin manning]

A very good start: http://www.freewebs.com/valvewizard/pentode.html

[Alexo]

I built a rig with variable plate, screen and cathode resistors and switchable cathode caps. Hooked up a couple meters to the plate and screen to make sure things didn't get too out of whack and adjusted till I found what my ears told me was the sweet spot.

Once I had my secret recipe, I plotted it on the charts - not what I would have chosen if I had worked the other way and done what looked best on paper, but it sure do sound good.

If you want it to sound good, you just have to tweak it till it does!